Efficient Training with Foresight: Multi-Token Auxiliary Supervision for Autoregressive Image Generation
Open paper
Benchmark
Model
Rank
Results
image-generation-on-imagenet-256x256
MTAR-B
–
FID: 4.50
image-generation-on-imagenet-256x256
MTAR-L
–
FID: 2.85
Rank counts only results with a code link.